Wednesday 11 January 2023

Legalised marijuana sales reduce the birth rate

The legalisation of marijuana in the US, where states have made medical marijuana sales legal, and then retail marijuana sales legal, and all at different times, has provided a wealth of possibilities for studying the effects of marijuana legalisation. From these studies, we have learned that medical marijuana sales may decrease harm from opiates, decrease violent crime, while legalised retail marijuana may increase house values (see also here), and lower crime rates, but may displace drug dealers into selling harder drugs. There is also evidence (from Europe) that legal access to marijuana reduces academic performance by students.

So, I was interested to read this recent working paper by Sarah Papich (University of California, Santa Barbara), which focuses on the effect of legalisation of marijuana sales on a completely different outcome - birth rates. The effect of marijuana use on the birth rate is theoretically ambiguous, because:

The medical literature suggests that marijuana use has two competing effects on fertility... First, marijuana use could lower the likelihood of pregnancy through effects on both men’s and women’s reproductive systems. Marijuana use is associated with lower sperm counts (Gundersen et al. 2015) and delayed ovulation (Bari et al. 2011), both of which make conception less likely. Second, marijuana use could lead to sexual behaviors that raise the likelihood of pregnancy. Using marijuana heightens the hedonic effect of sexual activity and diminishes the ability to think about long-term consequences of failing to use contraception. Marijuana use is associated with an increase in the amount of sexual activity (Sun and Eisenberg 2017) and a decrease in the likelihood of using contraception (Guo et al. 2002).

So, an increase in marijuana use could increase the birth rate, because people have more sex, and riskier sex, or it could decrease the birth rate, because marijuana users are less fertile. Papich uses a different-in-differences research design, which compares the difference in birth rates between states before marijuana is legalised, with the difference in birth rates between states after marijuana is legalised. Her key data come from the US National Vital Statistics System. Papich also distinguishes between the effects of legalising medical marijuana sales and legalising retail marijuana sales, while controlling for:

...shares of the total population by race, ethnicity, age, and education; unemployment rate; median household income; state cigarette tax; state beer tax; an indicator for whether the state has expanded Medicaid; indicators for abortion restrictions in the form of ambulatory surgical center laws, admitting privilege laws, and transfer agreement laws; an indicator for whether same-sex marriage is legal; the Medicaid eligibility threshold for pregnant women as a percentage of the federal poverty level; a WIC EBT indicator; and an indicator for whether marijuana has been decriminalised.

Papich finds that:

...days of marijuana use per month increase by 41% in response to RMLs and 23% in response to MMLs.

Ok, so people use more marijuana in response to medical marijuana laws (MMLs), and even more in response to recreational marijuana laws (RMLs). But what about birth rates? Papich then finds that:

...RMLs lead to a 2.78% decline in the average birth rate. This result provides evidence that marijuana’s physical effects, which suppress the likelihood of pregnancy conditional on sexual activity, have the dominant effect on fertility. Age heterogeneity analysis shows the largest decrease in the birth rate occurs among women 30-34, closely followed by women 35-39 and then by women 40-44. The birth rate in all three of these age groups declines by over 6%. This heterogeneity analysis suggests that women are having fewer total children in response to RMLs rather than delaying births...

I find that MMLs lead to a statistically insignificant decrease in birth rates.

Papich also looks at the effects on sexual activity, but those results are not as consistent, and not as convincing. However, she does provide some alternative evidence that sexual activity increases, or at least risky sexual activity increases:

RMLs are estimated to increase a state’s male gonorrhea cases by 6.1 cases per 100,000 population, a 5% increase from the mean. The effect of MMLs is statistically insignificant, with a positive point estimate.

The combination of those results suggests that legalising marijuana sales reduces birth rates, and the mechanism is likely to be through reduced fertility. In her conclusion, Papich notes some open questions that remain:

Data on contraceptive use before and after RMLs would provide insight into another mechanism through which marijuana legalization could affect fertility. Additional mechanisms, such as changes in the seriousness of romantic relationships when marijuana use increases and the effect of fewer people being imprisoned for marijuana possession, are promising areas for future research.

To those suggestions, I would add assessing whether the decrease in birth rates is primarily a result of reduced male fertility (since it appears that male marijuana use increased more than female marijuana use when recreational marijuana was legalised) or reduced female fertility. However, the headline result still holds - legalising marijuana sales reduces the birth rate.

[HT: Marginal Revolution, last year]

No comments:

Post a Comment